38
J. PIDANIČ, D. ČERMÁK, V. SCHEJBAL, GAIN ESTIMATION OF DOUBLY CURVED REFLECTOR ANTENNA
Gain Estimation of Doubly Curved Reflector Antenna
Jan PIDANIČ, Dušan ČERMÁK, Vladimír SCHEJBAL
Jan Perner Faculty of Transport, University of Pardubice, Studentská 95, 532 10, Pardubice, Czech Republic
jan.pidanic@upce.cz, dusan.cermak@upce.cz, vladimir.schejbal@upce.cz
Abstract. A simple formula of approximate gain estimation is verified for the doubly curved reflector antenna.
Numerical simulations using physical optics and experimental results of the shaped-beam doubly curved reflector
antenna are compared with the simple approximation of
gain. That approximation could be very valuable for system engineers to accurately estimate antenna gain and
coverage pattern and perform EMC calculations (estimations of interferences and susceptibilities) even for the
operation and out of operation frequency bands of shapedbeam antenna.
lyze not only radiation pattern of an antenna but various
related problems. In the early 1960’s, the set of programs
computing the shape and the radiation pattern of the doubly
curved reflector has been developed [6] - [8] and plenty of
Czech radar reflector antennas have been designed and
produced [9], [10]. The radiation characteristics of radar
RL-41 in the operation (2.7 to 2.9 GHz) and out of operation (3 to 10 GHz) frequency bands have been measured
and calculated using physical optics [11].
Keywords
Antenna gain, antenna beam width, doubly curved
reflector, estimation of antenna gain, electromagnetic
compatibility (EMC).
1. Introduction
Radar coverage and electromagnetic compatibility
(EMC) ask for accurate estimation of antenna gain, which
is the most important performance parameter of an antenna.
Sometimes, it is not possible to measure or calculate the
gain of an antenna. Many simple formulas are available for
gain estimating. The beam widths and directivity of a relatively large planar array or aperture antennas are related by
various well-known approximate equations [1] - [5]. That
approximation could be very valuable for system engineers
to accurately estimate antenna gain and perform EMC
calculations. As EMC calculations are very important for
radar antennas (estimations of interferences and susceptibilities), numerical simulations and experimental results are
compared with the simple approximation of gain for the
operation and out of operation frequency bands of shapedbeam antenna.
It is a common radar requirement to have a narrow
beam in one plane and a shaped beam in the other. The
shaped-beam doubly curved reflector antenna is a classical
reflector type, which was described in 1940’s. The far-field
radiation of antenna, shown in Fig. 1, can be calculated
using aperture method or physical optics [3] - [7]. Radar
antenna design can be done using suitable software. From
viewpoint of radar antenna systems, it is necessary to ana-
Fig. 1. Doubly curved reflector antenna.
The doubly curved reflector of RL-41 antenna is
shown in Fig. 2 with the secondary surveillance radar
(SSR) antenna on the top. The primary feed employs two
pyramidal horns, which create two independent beams. The
polarization can be continuously changed (horizontal –
elliptical – vertical). The lower and upper beams are shown
in Fig. 3. The angle θ is an elevation angle measured in the
vertical plane (symmetry plane) of the RL-41 antenna in
Fig. 2. The lower beam is designed as a nearly “pencil
beam” for lower elevations. The upper-beam is shaped and
its cosecant squared (cosec^2) vertical pattern is modified
for higher elevation angles.
2. Doubly Curved Reflector Antenna
The programs, used for computations of antenna
characteristics for doubly curved reflector antennas consisting of primary feed and reflector, have been coded,
debugged and gradually improved [7]. It computes far-field
rectangular components of primary feed electromagnetic
fields and therefore, it allows the electromagnetic field
computation at reflector surface. It is assumed that the
RADIOENGINEERING, VOL. 17, NO. 3, SEPTEMBER 2008
39
primary feed radiates the spherical waves. The program
offers several variants such as vertical or horizontal polarization calculations for horn antennas (pyramidal or conical) and radiation field calculations using aperture dimensions or approximations based on measured radiation
patterns in the main planes using the least squares method
with polynomials of order 6. Moreover, the primary feed
could be formed by one feed or an array of two or four
equally spaced radiators.
Fig. 2. The RL-41 antenna - doubly curved reflector with two
independent pyramidal horns and SSR antenna above.
The far field E in terms of the reflector complex current
density using physical optics [3], [7] is
E=−j
e − jkR
λR
∫ [n × (i
r
× E i )]e jk ir ⋅i R ⋅r d S
(1)
S
where R is the distance from the origin to the observation
position P (Fig. 1), n is the normal unit vector at the reflector surface, ir is the position vector from the origin to
the reflector, Ei is the incident electric field vector at the
reflector surface, determined by primary feed field calculations considering the relevant phases and the primary feed
gain, iR is the far-field observation position unit vector, r is
the distance from the origin to the reflector and S is the
reflector surface. The surface integral (1) can be evaluated
as a double integral of a function f (Y,Φ), where Y axis is
perpendicular to the symmetry plane and Φ is the angle
measured in the symmetry plane. Inner integrals F(Y) are
firstly evaluated and then outer integrals are calculated
over the variable Y (details can be found in [7]). If the gain
of primary feed is known, the gain of the whole antenna
could be calculated [7].
The reflector surface S and its contour cannot be determined analytically but the reflector central section is
calculated using geometrical optics and the surface S is
formed by parabolic ribs [3], [5] - [8]. Therefore, the input
data for surface S calculations are vectors of reflector central section and reflector contour. Moreover, Ei calculations
are very complicated. That means the far field (1) should
be calculated numerically. The crucial problem for physical
optics is the numerical integration as the integrand of the
radiation integral (1) oscillates rapidly for large reflectors.
It is therefore impossible to evaluate (1) analytically.
Moreover, integral (1) should be evaluated each time,
when the observation angle changes. Various methods for
physical optics have been suggested to increase the speed.
The very simple, accurate and powerful integration
methods have been proposed (details including the accuracy analyses can be found in [7], [8]), which would require substantially less RAM and computer time. The F(Y)
inner integrals over a variable Φ could be evaluated in
advance using the generalized trapezoidal method with 80
points. Then the Gaussian integration with respect to Y axis
is chosen. That allows numerical integrations for reflector
diameters DH greater than 100 wavelengths using just 8
F(Yi) values of inner integrals. Considering the first two or
three sidelobes 8 F(Yi) values could be only used for the
other observation positions with sufficient accuracy [7]. If
the primary feed movements and tolerances of reflector
surfaces are small, changes in amplitude can be neglected
and the phase changes could be only considered. That
means that computation (1) allows the determination of far
fields considering reflector surface deviations as well as
the feed movements. A slightly more complicated (more
time consuming) procedure should be used for bigger feed
movements.
Using physical optics the radiation patterns can be
only computed for angles near by the main beam (usually
the first two or three sidelobes), as the other sidelobes are
strongly affected by antenna tolerances [5], [8] and
neglecting of current at the shadow regions (that can be
calculated using another approximate methods such as
geometric theory of diffraction).
38
36
34
32
30
28
26
24
22
20
G [dB]
0
LB
5
UB
10
cosec^2
15
Θ [°]
20
Fig. 3. The RL-41 antenna - the lower and upper beams and
cosecant squared (cosec^2).
3. Experimental Results
The beam widths and directivity of a relatively large
planar array or aperture antennas are related by the wellknown simple approximate equation [1]:
D=
db
ϑ1ϑ2
(2)
40
J. PIDANIČ, D. ČERMÁK, V. SCHEJBAL, GAIN ESTIMATION OF DOUBLY CURVED REFLECTOR ANTENNA
where ϑ1 and ϑ2 are half-power beam widths in the principal planes of the elliptical beam at any scan angle, db is
constant (directivity-beam width product). The db values of
26 000 up to 52 525 could be considered for various cases.
The directivity of planar antennas has been calculated with
db = 32 400 and differences for various distributions have
been determined [1]. Therefore, the value of 32 400 could
be used. According to [1], the value of db = 26 000 could
be used for practical antennas (i.e. 1 dB difference against
32 400).
known that the horn quadratic phase error is directly proportional to frequency, and therefore for higher frequencies
the real feed beams are broader, the nulls between the main
lobe and the first sidelobe disappear and sidelobe levels are
raised [3] - [5]. Moreover, the horn phase characteristics
are changed. A possible radiation of higher modes creates
substantial changes of primary feed radiation patterns such
as maximum oscillations and sharper radiation patterns
[11]. Manufacturing tolerances (the surface imperfections)
change the optical path length from the feed to the reflector
aperture plane, and therefore phase errors are higher for
higher frequencies. Moreover the utilization of two horns,
which are not in focus, causes that the antenna is even
more tolerance sensitive.
D-1
D
G meas
G calcul
44
42
G [dB]
40
38
36
34
32
30
2
6
8
f [GHz] 10
Fig. 6. Measurements and calculations (using physical optics) of
gains and directivities for db =32 400 (D) and db =26 000
(D - 1) for lower beam.
Fig. 4. Measurements and calculations (using physical optics) of
ϑ1 half-power beam widths.
D-1
D
G meas
G calcul
Eq. (3)
38
37
36
G [dB]
Fig. 4 and 5 show the measurements and calculations
(using physical optics) of ϑ1 and ϑ2 half-power beam
widths for the lower beam (LB) and upper beam (UB) of
RL-41 antenna. The measurements and calculations (using
physical optics) of gains for RL-41 antenna are shown in
Fig. 6 and 7. That are compared with gain approximation
according to (2) for db = 32 400 (D) and db = 26 000 (D-1).
4
35
34
33
32
31
30
2
4
6
8
f [GHz] 10
Fig. 7. Measurements and calculations (using physical optics) of
gains and directivities for db =32 400 (D), db =26 000
(D – 1) for upper beam and Ds according to (3).
Fig. 5. Measurements of ϑ2 half-power beam widths.
The differences between measurements and calculations (using physical optics) for higher frequencies in Fig.
4 to 7 are given by simplifications. In this case, the rectangular waveguide was considered instead of real feed, i.e.
the rectangular aperture without any phase errors was used
for calculations. That creates higher differences for horn
with higher phase errors and tolerance problems. It is well
For the shaped beam (upper beam), it could be possible to
modify the simple approximation (2). Directivity can be
evaluated using the normalized radiation pattern F(θ,ϕ)
Ds =
≈
2π
π 2
0
−
∫ ∫π
ϕe ∫
π 2
−π 2
2
4π
F (θ , ϕ ) cos θ d θ d ϕ
2
4π
F (θ ) cos θ d θ
2
≈
4π
ϕ eθ e
(3)
RADIOENGINEERING, VOL. 17, NO. 3, SEPTEMBER 2008
41
where θe ,ϕe are the effective half-power beam widths of
the shaped beam. To estimate θe the integral could be
evaluated numerically considering the cosecant squared
vertical pattern of the upper beam. That approach could
offer slightly better estimation (see Fig. 7). However, that
is rather more complicated and cannot be generally used
(the radiation pattern is not usually known in advance).
Therefore, it is not generally possible to use (3) for simple
gain approximation.
[2] SCHEJBAL, V. Directivity of planar antennas. IEEE Antennas and
Propagation Magazine, 1999, vol. 41, no. 2, p. 60 - 62.
4. Conclusion
[7] KUPČÁK, D., SCHEJBAL, V. Calculating the radiation pattern of
doubly curved reflector antenna. Slaboproudý obzor, 1975, vol. 36,
no. 12, p. 567-571 (in Czech).
The paper presents the well-known simple formula
(2) of gain estimation. It can be seen that for the lower
beam, which can be roughly considered as a pencil beam,
the experimental results and calculations using physical
optics correspond with the simple formula of gain estimation according to (2) for the operation (2.7 to 2.9 GHz)
frequencies, when the limits of db = 32 400 and db = 26 000
are taken into account (i.e. 1 dB difference). For out of
operation (3 to 10 GHz) frequency bands, the differences
between (2) and calculations (using physical optics) are
nearly acceptable. The greater differences between experiments and calculations are explained above.
For the upper beam, which can be considered as a
shaped beam, the greater differences between experimental
results, calculations using physical optics and the simple
formula of gain estimation according to approximation (2)
exist. The approach (3) could offer slightly better estimation but it is rather more complicated and cannot be generally used. However, approximation (2) could be used as a
rough estimation.
The simple gain approximation could be very valuable for system engineers to accurately estimate antenna
gain and perform EMC calculations (estimations of interferences and susceptibilities). It seems that the approximations of D with db = 32 400 could be safely used as higher
limits for EMC calculations (interference and susceptibility
estimations). On the other hand, the approximations of D
with db = 26 000 could be used for radar equations or received power calculations of transmitting antenna.
Acknowledgements
The paper is supported by the Czech National
Institutional Research “Theory of Transport System” No.
MSM 0021627505.
References
[1] STUTZMAN, W. L. Estimating directivity and gain of antennas.
IEEE Antennas and Propagation Magazine, 1998, vol. 40, no. 4, p. 7
to 11.
[3] SILVER, S. Microwave Antenna Theory and Design. New York:
McGraw-Hill, 1949.
[4] KUHN, R. Mikrowellen Antennen. Berlin: VEB Verlag Technik,
1964 (in German).
[5] MILLIGAN, T. A. Modern Antenna Design. Hoboken: John Wiley
& Sons, 2005.
[6] KUPČÁK, D. Microwave antenna calculation using National Elliott
803 B computer. Radar Technology in Transport, 1965, Pardubice,
p. 20 - 34.
[8] SCHEJBAL, V., KUPČÁK, D. A survey of programs for calculating
microwave antennas with the aid of a computer. Slaboproudý obzor,
1976, vol. 37, no. 3, p. 117 – 122 (in Czech).
[9] SCHEJBAL, V. Czech radar technology. IEEE Trans. on Aerospace
and Electronics Systems, 1994, vol. 30, no. 1, p. 2 - 17.
[10] BEZOUŠEK, P., SCHEJBAL, V. Radar technology in the Czech
Republic. IEEE Aerospace and Electronic Systems Magazine, 2004,
vol. 19, no. 8, p. 27 – 34.
[11] SCHEJBAL, V., HONIG, J., KŘÍŽ, J. The radiation characteristics
of an antenna with a double-curvature reflector outside the working
frequencies range. Slaboproudý obzor, 1982, vol. 43, no. 12, p. 585
to 590 (in Czech).
About Authors...
Jan PIDANIČ was born in 1979. He received MS degree
from the Jan Perner Transport Faculty, University of Pardubice in 2005. He is interested in propagation of electromagnetic waves and in signal and data processing.
Dusan ČERMÁK was born in 1974. He received his MS
degree from the Czech Technical University in Prague in
2003 and his PhD degree from the University of Pardubice,
the Czech Republic in 2007. He has been with the University of Pardubice since 2003 as an assistant professor. He is
interested in simulations, measurements, radar antennas
and propagation of electromagnetic waves. He has published over 20 papers.
Vladimir SCHEJBAL graduated from the Czech Technical University, Prague in 1970. He received the PhD degree from the Slovak Academy of Science, Bratislava in
1980. He was with the Radio Research Institute Opocinek,
the Czech Republic (Antenna Department) from 1969 to
1993. From 1983 to 1986, he was on a leave with the
Higher Institute of Electronics (Microwave Department)
Beni Walid, Libya as a lecturer. He has been with the University of Pardubice, the Czech Republic since 1994, now
as a full professor and a head of the department. He is
interested in microwave antennas and propagation. He has
published over 100 papers. He is a senior IEEE member.